Middle-age spread 'seems to reduce dementia risk'

Genetics and stem cells

"Being overweight 'reduces dementia risk'," BBC News reports. The story comes from a cohort study of nearly 2 million UK adults aged over 40. It showed that being overweight or obese was linked to a lower risk of dementia…

"Being overweight 'reduces dementia risk'," BBC News reports. The story comes from a cohort study of nearly 2 million UK adults aged over 40. It showed that being overweight or obese was linked to a lower risk of dementia up to 20 years later, compared with people who were a healthy weight. Underweight people were at a higher risk of dementia.

This result is surprising as it contradicts the current consensus of opinion, including the advice on this website, that obesity may be a risk factor for some types of dementia.

In the best scientific tradition, this study raises more questions than it answers. But it is important not to overlook the many serious health risks associated with obesity, such as heart disease and diabetes.

As one of the key authors, Dr Qizilbash, rightly says, the findings are "not an excuse to pile on the pounds or binge on Easter eggs … You can't walk away and think it's OK to be overweight or obese. Even if there is a protective effect, you may not live long enough to get the benefits".

In conclusion, a single study is unlikely to lead to a change in clinical guidelines, but it is likely to prompt further research into the issue. 

Where did the story come from?

The study was carried out by researchers from the London School of Hygiene and Tropical Medicine, London, and OXON Epidemiology; a London/Madrid-based clinical research company.

The study reports no funding for the work and the authors declare no conflicts of interest.

It was published in the peer-reviewed medical journal, The Lancet Diabetes & Endocrinology.

Generally, the media reported the story accurately and responsibly, taking a range of angles. The Daily Telegraph outlined how "a middle age spread may protect against dementia"; The Guardian said "underweight people face significantly higher risk"; while The Independent went with a lack of risk angle, saying that, "being overweight may not increase dementia risk" as previously thought. All accurately reflect the results of the underlying study.

Much of the news outlined how these findings contradict previous research, but may be more reliable because the study was bigger and more robust. Most also cautioned against taking this to mean that being overweight or obese is somehow good for your health, and said the link between dementia and obesity was an open case, needing more research to find out what's going on.

What kind of research was this?

This was a retrospective cohort study looking at body mass index (BMI) and dementia, using information from UK GP records.

BMI is a measure of weight and height. The main four BMI categories – underweight, healthy weight, overweight and obese – are based on whether your weight is likely to affect your health.

The healthy weight category means your weight isn't likely to affect your health, whereas the overweight category means your weight is likely to increase your chance of death and disease. This is the same for the underweight category. Obese people are more likely to suffer death and disease than people who are overweight. 

This type of study cannot prove cause and effect, but can give us an idea of possible links. One of the disadvantages of using existing GP records is you can only use the information that has already been collected. This might not include all the information you would want to collect as a researcher, such as changes in body weight, physical activity levels, diet, and other lifestyle factors.  

What did the research involve?

The researchers analysed more than 1.9 million UK GP records to see whether BMI was linked to a recorded diagnosis of dementia.

The cohort of people analysed were all over 40, had no previous diagnosis of dementia, and had to have a BMI measure recorded in their GP notes between 1992 and 2007. Everyone else was excluded.

Eligible medical records were reviewed to see if people went on to develop dementia, changed GP practice, or died up to July 2013. The average time elapsed between the single BMI measurement and any of these events was nine years. Some had records spanning 20 years. 

The team split the people into standard BMI categories and calculated their relative risk of developing dementia. The categories were:

  • underweight: BMI less than 20kg/m2
  • healthy weight: BMI 20 to less than 25kg/m2
  • overweight: BMI 25 to 30kg/m2
  • obese: BMI greater than 30kg/m2, actually divided into three subcategories of obesity: class I, II and III

The analysis adjusted for a range of known confounders already recorded in the GP records, including:

  • age
  • gender
  • smoking
  • alcohol consumption
  • history of heart attack, stroke or diabetes
  • recent use of statins or drugs to treat high blood pressure

What were the basic results?

Dementia affected 45,507 people, just over 2 out of every 100 taking part (crude prevalence 2.32%).

Compared with people of a healthy weight, underweight people had a 34% higher risk of dementia (rate ratio [RR] 1.34 95% confidence interval [CI] 1.30 to 1.39).

Compared with people of a healthy weight, overweight people had a 19% lower risk of dementia (RR 0.81, 95% CI 0.79 to 0.83). The incidence of dementia continued to fall marginally for every increasing BMI category, with very obese people (BMI greater than 40kg/m2) having a 33% lower dementia risk than people of a healthy weight (RR 0.67, 95% CI 0.60 to 0.74).

These patterns stayed stable throughout two decades of follow-up, after adjustment for potential confounders and allowance for the J-shape association of BMI with mortality. 

How did the researchers interpret the results?

The research team says: "Our study shows a substantial increase in the risk of dementia over two decades in people who are underweight in mid-life and late-life.

"Our findings contradict previous suggestions that obese people in mid-life have a higher subsequent risk of dementia. The reasons for and public health consequences of these findings need further investigation." 


This cohort study of more than 1.9 million UK adults aged over 40 links being overweight or obese to a lower risk of dementia, compared with healthy weight people. Underweight people were at a higher risk of dementia.

The study has many strengths, such as its large size and applicability to the UK. However, the authors note their results buck the trend of other research, which found being overweight or obese was linked to an increase risk. They suggest their study is probably more reliable than the past ones as they were smaller.

They aren't quite sure what this means, and say: "The reasons for and public health consequences of these findings need further investigation."

It's important to realise that this finding doesn't mean that gaining weight will somehow protect you against dementia. Many dietary, environmental and genetic factors are likely to influence both BMI and dementia, so the relationship is complex.

However, we do know that being overweight or obese is bad for your health. The same is true for people who are underweight as they are not getting the nutrients their body needs, which may be one of the reasons why they were found to have an increased risk of dementia in this study.

Dr Liz Couthard, Consultant Senior Lecturer in Dementia Neurology at the University of Bristol, said: "We do know that obesity carries many other risks, including high blood pressure, heart disease, diabetes and increased rates of some types of cancer. So maintaining a healthy weight is recommended."

However, there are limitations to bear in mind with this study that may have affected the findings to some degree.

Selection bias

First is the possibility of selection bias. Around half (48%) of eligible people did not have a BMI record, so were excluded from the study. A further third (31%) with BMI records were excluded for not having at least 12 months of previous health records. The study team were aware of this, saying: "If BMI is more likely to be measured in people with comorbidities than in healthy people, which might in turn be associated with dementia risk, then some bias is possible." But they went on to say this is unlikely.


Residual confounding is also a possibility. The researchers had to use variables collected in the GP records, which didn't cover everything they would have wanted. For example, they adjusted for anti-hypertensive medicines and statins but not for blood pressure and blood lipid values, which, they say, do affect the associations of BMI with heart attack and stroke.

Unavailable data

Other unavailable potential confounders, such as physical activity level, socioeconomic status and ethnic origin, might also have influenced the recorded association between BMI and dementia. We can't say to what extent.

Maintaining a healthy weight is recommended to reduce the risk of heart disease, diabetes and some cancers. This study suggests the benefits of this may not extend to reducing the risk of dementia, but the relationship is likely to be complex and is not yet fully understood.

Article Metadata Date Published: Do, 23 Nov 2017
Author: Zana Technologies GmbH
NHS Choices